First, please follow NVIDIA's specific instructions to setup your Jetson Development Kit (see Jetson Development Kit User Guides). Once you have installed the NVIDIA Jetpack on your Jetson Development Kit, make sure all system libraries are up-to-date. In a terminal, run:
sudo apt-get update
sudo apt-get upgrade
Lastly, please test that CUDA is installed properly by running: nvcc --version
. The output should say the version of CUDA installed on your Jetson.
We highly recommend installing DeepLabCut-live in a virtual environment. Please run the following command to install system dependencies needed to run python, to create virtual environments, and to run tensorflow:
sudo apt-get update
sudo apt-get install libhdf5-serial-dev \
hdf5-tools \
libhdf5-dev \
zlib1g-dev \
zip \
libjpeg8-dev \
liblapack-dev \
libblas-dev \
gfortran \
python3-pip \
python3-venv \
python3-tk \
curl
Next, create a virtual environment called dlc-live
, activate the dlc-live
environment, and update it's package manger:
python3 -m venv dlc-live
source dlc-live/bin/activate
pip install -U pip testresources setuptools
First, install python dependencies to run tensorflow (from NVIDIA instructions to install tensorflow on Jetson platforms). This may take ~15-30 minutes.
pip3 install numpy==1.16.1 \
future==0.17.1 \
mock==3.0.5 \
h5py==2.9.0 \
keras_preprocessing==1.0.5 \
keras_applications==1.0.8 \
gast==0.2.2 \
futures \
protobuf \
pybind11
Next, install tensorflow 1.x. This command will depend on the version of Jetpack you are using. If you are uncertain, please refer to NVIDIA's instructions. To install tensorflow 1.x on the latest version of NVIDIA Jetpack (version 4.4 as of 8/2/2020), please the command below. This step will also take 15-30 mins.
pip3 install --pre --extra-index-url https://developer.download.nvidia.com/compute/redist/jp/v44 'tensorflow<2'
Lastly, copy the opencv-python bindings into your virtual environment:
cp -r /usr/lib/python3.6/dist-packages ~/dlc-live/lib/python3.6/dist-packages
Finally, please install DeepLabCut-live from PyPi (this will take 3-5 mins), then test the installation:
pip install deeplabcut-live
dlc-live-test
If installed properly, this script will i) download the full_dog model from the DeepLabCut Model Zoo, ii) download a short video clip of a dog, and iii) run inference while displaying keypoints.